In recent years there has been an increasing interest in alcoholic beverages with a reduced alcohol content. As a result, numerous low-alcohol beers have appeared on the market, and there have been similar efforts to produce "light" wines. When first introduced, light wines accounted for about 10% of the total $7 billion (550 million gallon) U.S. wine market. Largely because of problems with the presently available methods for producing light wines and resultant poor quality, however, sales of such wines have fallen to 3% of total wine sales.
Production of low-alcohol beers has not proven to be difficult since the chemistry of beer production is relatively simple compared to that of wine, and beer can tolerate rougher treatment. Thus, light beer may be produced by boiling regular beer for a number of hours to drive off much of the alcohol. Hoyrup, "Beer," in Kirk-Othmer Encyclopedia of Chemical Technology, Vol. 3, pp. 692-735 (3rd Ed. 1978). Such protracted heating of wine would degrade many of the constituents that contribute to its flavor, color and bouquet. With beer, flavor that is lost by boiling may be restored to some degree by the addition of aroma substances recovered from yeast (German Pat. No. 1,767,040), but there is no such simple remedy for the restoration of flavor to thermally damaged wine.
The boiling of beer to remove alcohol also causes a substantial loss of water. That poses no problem for beer because it can simply be reconstituted by the addition of replacement water. Dilution of wine or distilled spirits with make-up water, however, is restricted or prohibited by the U.S. Bureau of Alcohol, Tobacco, and Firearms. 49 Fed. Reg. 37510-37530, (Sept. 24, 1984). Other methods for producing low-alcohol beer that also cause the removal of water, such as vacuum distillation and reverse osmosis, are not applicable to wine because of this prohibition. Where the ethanol content of distilled spirits such as whiskeys is reduced by dilution with water, the product must be labelled as "diluted," and this is undesirable from a marketing standpoint.
Yet another process for reducing the alcohol content of beer involves the dialysis of beer using cellulosic dialysis membranes. Moonen, H. and H. J. Niefind, "Alcohol Reduction in Beer by Means of Dialysis," Desalination, Vol. 41, p. 327-335 (1982). In this process, beer and water flow countercurrently on opposite sides of a dialysis membrane, and alcohol from the beer stream diffuses across the membrane to be swept away at low concentration in the water (or "dialysate") stream. Unfortunately as regards the production of light wines, however, simple dialysis of wine against water across relatively non-selective dialysis membranes would not only remove ethanol but would also remove many low- to middle-molecular-weight solutes essential to the flavor and bouquet of a wine. Additionally, an osmotically-driven diffusive flux of water from the dialysate side to the wine side of the membrane could be expected to occur, and such a water flux would constitute a form of dilution or "reconstitution" of the beverage.
Efforts have been made to produce low-ethanol wine through the use of low-sugar grapes or by premature curtailment of the fermentation process, but the complexity and quality of the product are adversely affected.
A better process has been described by Boucher in U.S. Pat. No. 4,405,652, in which fully mature wine is subjected to flash evaporation under vacuum in a heated centrifugal evaporator. Because the wine is spun into a thin film, from which rapid ethanol evaporation takes place, and because the evaporation process is conducted under a vacuum of 27-29 inches Hg, degradation of the wine is reduced. Nevertheless, there may be some thermal destruction since the wine is heated to 98.degree.-100.degree. F., and in fact wine processed in this manner is said to exhibit a characteristic and undesirable "burnt" taste.
Additionally, other volatile components that may be important to the flavor, color or bouquet of the wine may be lost during the evaporation phase of the process. Finally, the product claimed by Boucher is only modestly reduced in ethanol content to about 7.1 volume percent from an original wine alcohol content of 10.8 volume percent.
Pressure-driven membrane processes such as reverse osmosis, ultrafiltration, and pervaporation, even if they were carried out at low temperatures (under relatively high vacuum in the case of pervaporation) and with suitably ethanol-permselective membranes, might in fact be sufficiently non-destructive to wine, but such processes would also remove excessive quantities of water and concentrate the alcoholic beverage in the process. As previously noted, lost water cannot simply be replaced in wine with water from an external source. Additionally, the high osmotic pressure differences that would be encountered in the production of a high-purity ethanol permeate from such relatively ethanol-poor mixtures as wine (10-12 volume percent ethanol) and whiskeys (40 volume percent ethanol) would make reverse osmosis both technically and economically unfeasible. Indeed, at the present time reverse osmosis is more typically applied to the task of concentrating ethanol in aqueous mixtures by removal of water therefrom than it is applied to the selective removal of ethanol from such mixtures.
Conventional solvent extraction technology has long been applied to the recovery of ethanol from aqueous solutions in industry. Scheibel, E. G., "Dehydration of Ethyl Alcohol by Fractional Liquid Extraction," Industrial & Engineering Chemistry, Vol. 42, p. 1497-1508 (1950). This technology, however, is not directly applicable to the production of low-alcohol wines or other beverages. There would invariably be excessive solubility of the extraction solvent in the wine and, hence, contamination. Emulsification and physical entrainment might also occur. Hartline, F. F., "Lowering the Cost of Alcohol," Science, Vol. 206, pp. 41-42 (1979). Furthermore, with most extraction solvents it would be expected that numerous other organic constituents of the wine would be coextracted with the ethanol, thereby creating a wholly unacceptable product.
Membrane solvent extraction, in which a membrane is interposed between a solvent containing a solute to be extracted and a second, immiscible extraction solvent, prevents the solvent entrainment and emulsion formation problems inherent to conventional solvent extraction technology. For example, Kim, in U.S. Pat. No. 4,443,414, used a microporous membrane to extract molybdenum from solutions containing molybdenum and other mineral ions. Lee et al., in U.S. Pat. No. 3,956,112, described a membrane solvent extraction system for general application based upon the use of a non-porous membrane. The membrane was solvent swollen, so that one of two substantially immiscible liquids which the membrane separated caused the membrane to swell, forming an intermediary zone through which diffusion of solute material could occur. Ho et al., in U.S. Pat. No. 3,957,504, used an ion-exchange membrane in the manner of Lee et al., to recover metal ions from an aqueous solution.
Because the above membrane solvent extraction systems involve the use of solvent-swollen membranes, they do not prevent the molecular diffusion of dissolved solvent into the aqueous phase. Furthermore, the membranes of the prior-art systems show no permselectivity for the solutes to be removed. Instead, any selectivity observed is due to the choice of the extraction solvent or to the inclusion of chelating agents in the solvent that are selective for the metal ions that are to be extracted. Finally, the organic extraction solvents employed by Ho et al., and by Lee et al., would be quite unsuitable for the production of beverages such as low-alcohol wines, distilled spirits, and beers for the reason that even minor amounts of these solvents, when dissolved in the aqueous phase, would represent sometimes toxic and invariably unacceptable contaminants or adulterants in the beverage.
What is needed for the removal of ethanol and other low-molecular weight organic solutes from aqueous solutions of these solutes--and in particular, from alcoholic beverages--is a process with the following characteristics:
Ethanol should be removed as selectively as possible, i.e. with minimal simultaneous removal of water. PA1 Ethanol should be removed in such a way that addition of water to the product (i.e., "reconstitution") is avoided. Neither deliberate addition of water (i.e., dilution) nor inadvertent addition of water (i.e., by direct osmosis or ultrafiltration of water into the beverage) should take place. PA1 Most organic compounds present in the beverage other than ethanol should be retained in the beverage during ethanol removal to the greatest degree possible. PA1 The extraction fluid that serves the purpose of receiving the ethanol extracted from the alcoholic beverage and subsequently carrying it away for further processing must either be such that its presence in the beverage is permissible at the concentration levels at which it will be found, or this fluid must be physically prevented from entering and thereby contaminating the alcoholic beverage that is being processed.
The invention described herein satisfies these criteria.